AI image content policy

Forbidden prompts and uploads

You may not prompt the AI to generate, nor upload as scene material, any of the following:

• Sexual content involving minors. Zero tolerance. Any attempt to generate or upload such content is reported to authorities and triggers immediate, permanent account termination without refund.
• Sexual or pornographic content. Including nudity, suggestive imagery, fetish material, and content depicting non-consensual sexual scenarios.
• Hate symbols and harassment. Imagery glorifying violence against protected groups, hate-group iconography, or content that targets, dehumanises, or harasses individuals.
• Graphic violence, gore, or animal cruelty. Including dismemberment, severe injury, mutilation, or torture imagery.
• Self-harm, suicide, or eating-disorder content. Including content that promotes, glorifies, or instructs on these topics.
• Illegal activity. Drug paraphernalia, weapons manufacturing, terror imagery, or promotion of crimes against persons or property.
• Photorealistic depictions of real peoplewithout their consent, including living public figures, politicians, and celebrities. This includes attempts to generate “a photo of [Person Name]”, deepfake-style portraits, or content that could mislead viewers about who they are seeing. Stylised / clearly-fictional characters that don't resemble specific individuals are fine.
• Copyrighted characters and brands.Disney characters, Pokémon, Marvel / DC superheroes, brand logos, mascots, video-game IP, and other trademarked or copyrighted material. The model will refuse most such prompts; the ones that slip through are still your liability, not ours or OpenAI's.
• Election manipulation and political deception. Imagery designed to mislead voters, fabricated quotes attributed to politicians, or deepfakes of public officials.
• Medical, legal, or financial advice imagery presented as authoritative without proper credentials. Pseudoscientific health claims, unverified medication dosages, and similar imagery harm consumers.
• Personal information and doxxing. Imagery containing real addresses, phone numbers, government IDs, financial details, or other identifying data of private individuals.

Sensitive content (allowed with care)

The following are not outright forbidden but are flagged by OpenAI's safety systems and often refused. Generation may fail; if it does, your credits are refunded. Examples:

• · Children depicted in non-sexual contexts (educational content is fine; close framing of unaccompanied minors is often refused).
• · Religious imagery — generally allowed; depictions of religious figures may be refused depending on context.
• · Historical violence (war, slavery, atrocities) — allowed for educational framing; gratuitous depictions are refused.
• · Mild substance use — allowed in artistic / lifestyle contexts; instructional imagery for production is refused.
• · Public figures in clearly satirical or editorial contexts — sometimes allowed at OpenAI's discretion.

If a render is refused upstream, NVE auto-refunds the per-image credit charge. The orchestration / planning credits already debited are not refunded — those covered work that ran successfully.

Your responsibility

• Prompts are yours.NVE forwards your prompts and briefs to OpenAI / Anthropic verbatim (with our scene-style suffix appended). We do not pre-screen prompts beyond the validation noted above; we rely on the upstream provider's safety classifiers. You are the author and owner of the prompt.
• Outputs are yours.You hold the rights to images NVE generates from your prompts, subject to OpenAI's own terms. NVE does not claim copyright or licensing rights to your generated content.
• Final-publish responsibility. Before you upload the resulting MP4, Shorts, or images to YouTube / Spotify / Apple / your own platform, you must review for accuracy, IP / brand compliance, and appropriateness for your audience. Automated moderation is not editorial review.
• Children's content (kids audience).If you produce content for children under 13, follow the relevant platform's rules (YouTube's Made-for-Kids designation, COPPA, GDPR-K). Avoid imagery that mimics copyrighted children's IP. Disclose AI-generated visuals in the description per platform policy.

Enforcement

• First-line enforcement is OpenAI's. DALL·E refuses prompts that hit its classifiers. Refused renders return your image-credit cost. The intake / planning phase already paid is not refunded.
• Pattern-of-abuse enforcement is ours. If your account repeatedly attempts to generate forbidden content (CSAM, non-consensual real-person deepfakes, hate imagery, etc.), we will suspend the account, void remaining credits, and where required by law forward records to authorities.
• Uploaded scene imagesare treated identically. Don't upload anything you would not be willing to defend in writing — copyright owner, parent/guardian releases, model releases, the lot. We do not store releases on your behalf and you remain solely responsible.

Reporting violations

If you believe content generated through NVE infringes your rights or violates the policies above, email nve@berta.one with the project ID, the specific concern, and any supporting evidence. We respond within five business days. CSAM reports are escalated immediately to the relevant authorities and the responsible account is terminated regardless of the report's outcome.